🚀 go-pugleaf

I just took a look at this thread so let me make a few comments.  The
first thing to understand is that supercite *by definition* doesn't
really know anything about VM (or GNUS, RMAIL, or any other reader it
can be connect to).  There is a well defined interface which SC
expects and which VM meets as distributed (yay Kyle!).  Other readers,
such as GNUS need certain functions to be overloaded in order to meet
this interface.  See the SC2.2 texinfo document for more information
on this interface.

SC is only called on a region of the reply buffer; ie point and mark
delineate the area that SC will cite.  The mail headers which SC will
use to extract information must be at the top of this region, and this
is the only information that SC parses before it starts to do its
thing.  So looking at the "From " line (no colon) isn't an option. SC
never sees the original mail or news article that might contains this
line.

SC expects that *all* mail headers be inserted into the reply buffer
region.  Its SC's job to remove what's not wanted based on the user's
preferences (there are sc-* variables for controlling this). I don't
know why your VM would not be including the From: line in the reply,
but it should.  You can take a look at the mail headers inserted by
toggling sc-nuke-mail-headers-p to nil (you may also want to setq
sc-auto-fill-region-p to nil, then once the article is cited, hit undo
once).  If the From: line's not there, its a problem with VM, but
you're right, SC won't be able to figure out things like the author's
name and email address if there's no From: line.  SC2.2 should be a
bit saner than 2.1 in writing the header when there's no From: line.

Also, I don't like the idea of SC parsing the X-VM-v5-Data for
information.  SC is supposed to be a general package, not specific to
VM, and I don't want to have to track any future changes to this that
Kyle may implement for VM enhancements.

First figure out why there's no From: line in your reply buffer, and
we'll go from there.

-Barry

PS. I personally use VM 5.32 (and GNUS 3.14) with SC 2.2, so I *know*
it works :-)